Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-04-14 Thread Dima Pasechnik
On Tue, Apr 14, 2020 at 9:47 PM Antonio Rojas wrote: > > That's osx, not Arch oops, sorry for noise. > > El martes, 14 de abril de 2020, 15:38:27 (UTC+2), Dima Pasechnik escribió: >> >> On Tue, Apr 14, 2020 at 9:28 PM Antonio Rojas wrote: >> > >> > I'm missing some context here... flint does ce

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-04-14 Thread Antonio Rojas
That's osx, not Arch El martes, 14 de abril de 2020, 15:38:27 (UTC+2), Dima Pasechnik escribió: > > On Tue, Apr 14, 2020 at 9:28 PM Antonio Rojas > wrote: > > > > I'm missing some context here... flint does certainly link to ntl > already, i don't see how it could work otherwise. > > David ju

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-04-14 Thread Dima Pasechnik
On Tue, Apr 14, 2020 at 9:28 PM Antonio Rojas wrote: > > I'm missing some context here... flint does certainly link to ntl already, i > don't see how it could work otherwise. David just posted a message here claiming this not to be the case on his Arch installation. Perhaps he has an odd version

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-04-14 Thread Antonio Rojas
I'm missing some context here... flint does certainly link to ntl already, i don't see how it could work otherwise. El martes, 14 de abril de 2020, 15:23:08 (UTC+2), Dima Pasechnik escribió: > > Hi Antonio, > can Arch's Flint be fixed so that it links against NTL, like in Sage? > Then we won't

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-04-14 Thread Dima Pasechnik
Hi Antonio, can Arch's Flint be fixed so that it links against NTL, like in Sage? Then we won't have to build Flint (and Arb) on Arch... On Sat, Jan 11, 2020 at 3:53 PM Antonio Rojas wrote: > > Build fails at fflas-ffpack, log attached. Seems related to FS#27870 > > -- > You received this messa

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-01-23 Thread Isuru Fernando
It depends on the fortran name mangling. For gfortran, one symbol would be dgeqrf_. Isuru On Thu, Jan 23, 2020 at 12:05 PM Dima Pasechnik wrote: > On Thu, Jan 23, 2020 at 5:36 PM Antonio Rojas wrote: > > > > > > > > El jueves, 23 de enero de 2020, 18:11:51 (UTC+1), Dima Pasechnik > escribió: >

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-01-23 Thread Dima Pasechnik
On Thu, Jan 23, 2020 at 5:36 PM Antonio Rojas wrote: > > > > El jueves, 23 de enero de 2020, 18:11:51 (UTC+1), Dima Pasechnik escribió: >> >> On Wed, Jan 22, 2020 at 5:20 PM Isuru Fernando wrote: >> > >> > [root@50586643ff22 /]# pkg-config --modversion blas >> > 0.3.7 >> > [root@50586643ff22 /]#

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-01-23 Thread Antonio Rojas
El jueves, 23 de enero de 2020, 18:11:51 (UTC+1), Dima Pasechnik escribió: > > On Wed, Jan 22, 2020 at 5:20 PM Isuru Fernando > wrote: > > > > [root@50586643ff22 /]# pkg-config --modversion blas > > 0.3.7 > > [root@50586643ff22 /]# pkg-config --modversion cblas > > 3.9.0 > > [root@50586643

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-01-23 Thread Isuru Fernando
> One way to do this is to test openblas for a symbol that is on Arch in cblas, not in openblas, I'm not sure I understand. Are you asking for a symbol in libcblas.so that is not in an libopenblas.so compiled with CBLAS? > Could someone figure out these symbols, or give me access to an Arch box?

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-01-23 Thread Dima Pasechnik
On Thu, Jan 23, 2020 at 5:11 PM Dima Pasechnik wrote: > > On Wed, Jan 22, 2020 at 5:20 PM Isuru Fernando wrote: > > > > [root@50586643ff22 /]# pkg-config --modversion blas > > 0.3.7 > > [root@50586643ff22 /]# pkg-config --modversion cblas > > 3.9.0 > > [root@50586643ff22 /]# pkg-config --modversi

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-01-23 Thread Dima Pasechnik
On Wed, Jan 22, 2020 at 5:20 PM Isuru Fernando wrote: > > [root@50586643ff22 /]# pkg-config --modversion blas > 0.3.7 > [root@50586643ff22 /]# pkg-config --modversion cblas > 3.9.0 > [root@50586643ff22 /]# pkg-config --modversion lapack > 3.9.0 > [root@50586643ff22 /]# pkg-config --modversion open

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-01-22 Thread Isuru Fernando
[root@50586643ff22 /]# pkg-config --modversion blas 0.3.7 [root@50586643ff22 /]# pkg-config --modversion cblas 3.9.0 [root@50586643ff22 /]# pkg-config --modversion lapack 3.9.0 [root@50586643ff22 /]# pkg-config --modversion openblas 0.3.7 On Wed, Jan 22, 2020 at 11:18 AM Dima Pasechnik wrote: >

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-01-22 Thread Dima Pasechnik
On Wed, Jan 22, 2020 at 5:03 PM Isuru Fernando wrote: > > > On Arch, does cblas.pc exist? > > Yes. Thanks. I wonder what on Arch is the output of pkg-config --modversion X for X in [cblas,blas,lapack,openblas] > > > this is system-dependent. > > Yes, that's why we need to have fallbacks. > >

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-01-22 Thread Isuru Fernando
> On Arch, does cblas.pc exist? Yes. > this is system-dependent. Yes, that's why we need to have fallbacks. Isuru On Wed, Jan 22, 2020 at 11:00 AM Dima Pasechnik wrote: > On Wed, Jan 22, 2020 at 4:17 PM Isuru Fernando wrote: > > > > > maybe they split openblas.so into separate sections for

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-01-22 Thread Dima Pasechnik
On Wed, Jan 22, 2020 at 4:17 PM Isuru Fernando wrote: > > > maybe they split openblas.so into separate sections for an unclear to me > > reason, but most applications I know uses cblas, blas and lapack, so they > > need to link them all anyway. > > True, but in sage you are copying `openblas.pc`

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-01-22 Thread Isuru Fernando
> maybe they split openblas.so into separate sections for an unclear to me reason, but most applications I know uses cblas, blas and lapack, so they need to link them all anyway. True, but in sage you are copying `openblas.pc` to `cblas.pc` and `lapack.pc` which is wrong. `openblas.pc` should only

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-01-22 Thread Isuru Fernando
> errors we see here, regarding absence of certain openmp functions in cblas, seem to indicate that cblas on Arch does not come from openblas. perhaps what we see are reference cblas and blas from openblas installed at the same time, in error. Can you please explain more? As I said earlier, it doe

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-01-22 Thread Dima Pasechnik
On Wed, 22 Jan 2020, 15:59 Isuru Fernando, wrote: > > But Arch does not do this, it instead provides dog-slow CBLAS and > LAPACK, built > from another (reference, i.e. no assembler, no optimisation) > implementation. > > This is simply not true. CBLAS's performance does not depend on which > impl

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-01-22 Thread Isuru Fernando
There's also a 3rd bug which is in fflas-ffpack which assumes that if there's no cblas, then it's not openblas. This can be fixed by copying the lines https://github.com/linbox-team/fflas-ffpack/blob/2d3af1a5bec51983b5a896fae12e60914d9bdc4d/fflas-ffpack/config-blas.h#L315-L317 to Line 311 above the

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-01-22 Thread Isuru Fernando
> But Arch does not do this, it instead provides dog-slow CBLAS and LAPACK, built from another (reference, i.e. no assembler, no optimisation) implementation. This is simply not true. CBLAS's performance does not depend on which implementation it comes from. CBLAS's performance depends on the unde

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-01-22 Thread Matthias Koeppe
On Wednesday, January 22, 2020 at 10:27:02 AM UTC-5, Dima Pasechnik wrote: > > > a meaningful resolution of this would be to test openblas for lapack and > cblas capacities. > > this is relatively easy, and would result in openblas on Arch being built > by Sage. > +1 on doing this. -- You re

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-01-22 Thread Dima Pasechnik
On Wed, 22 Jan 2020, 15:12 Vincent Delecroix, <20100.delecr...@gmail.com> wrote: > Le 22/01/2020 à 15:41, Dima Pasechnik a écrit : > > On Wed, Jan 22, 2020 at 2:31 PM Vincent Delecroix > > <20100.delecr...@gmail.com> wrote: > >> > >> I had the same annoying trouble... Is it really a bug in arch? I

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-01-22 Thread Vincent Delecroix
Le 22/01/2020 à 15:41, Dima Pasechnik a écrit : On Wed, Jan 22, 2020 at 2:31 PM Vincent Delecroix <20100.delecr...@gmail.com> wrote: I had the same annoying trouble... Is it really a bug in arch? I don't find anywhere where libopenblas is supposed to provide the LAPACK interface. To me it is ju

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-01-22 Thread Dima Pasechnik
On Wed, Jan 22, 2020 at 2:31 PM Vincent Delecroix <20100.delecr...@gmail.com> wrote: > > I had the same annoying trouble... Is it really a bug in arch? I don't > find anywhere where libopenblas is supposed to provide the LAPACK > interface. To me it is just true for SageMath and false for Archlinux

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-01-22 Thread Vincent Delecroix
Moreover, SageMath does work on archlinux perfectly well when installed from the package manager. Antonio, do you have some tweaks to deal with the blas/lapack settings when compiling for archlinux? Le 22/01/2020 à 16:31, Vincent Delecroix a écrit : I had the same annoying trouble... Is it reall

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-01-22 Thread Vincent Delecroix
I had the same annoying trouble... Is it really a bug in arch? I don't find anywhere where libopenblas is supposed to provide the LAPACK interface. To me it is just true for SageMath and false for Archlinux. Le 13/01/2020 à 11:44, Dima Pasechnik a écrit : On Sat, Jan 11, 2020 at 11:35 AM arojas

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-01-13 Thread Dima Pasechnik
On Sat, Jan 11, 2020 at 11:35 AM arojas wrote: > > El sábado, 11 de enero de 2020, 11:23:32 (UTC+1), Dima Pasechnik escribió: >> >> >> >> On Sat, 11 Jan 2020, 10:19 arojas, wrote: >>> >>> I suspect this may be due to our openblas package only providing libblas.so >>> (not libcblas.so or liblapac

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-01-11 Thread arojas
El sábado, 11 de enero de 2020, 11:23:32 (UTC+1), Dima Pasechnik escribió: > > > > On Sat, 11 Jan 2020, 10:19 arojas, > wrote: > >> I suspect this may be due to our openblas package only providing >> libblas.so (not libcblas.so or liblapack.so, which in our case come from >> the netlib blas). >>

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-01-11 Thread Dima Pasechnik
On Sat, 11 Jan 2020, 10:19 arojas, wrote: > I suspect this may be due to our openblas package only providing > libblas.so (not libcblas.so or liblapack.so, which in our case come from > the netlib blas). > Are you saying that your libopenblas.so does not provide everything that is in openblas by

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-01-11 Thread arojas
I suspect this may be due to our openblas package only providing libblas.so (not libcblas.so or liblapack.so, which in our case come from the netlib blas). El sábado, 11 de enero de 2020, 10:38:07 (UTC+1), Dima Pasechnik escribió: > > > > On Sat, 11 Jan 2020, 09:21 Antonio Rojas, > > wrote: > >

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-01-11 Thread Dima Pasechnik
On Sat, 11 Jan 2020, 09:21 Antonio Rojas, wrote: > openblas is not being built in Sage, the system one is detected and used > after FS#27870. FWIW, building our distro fflas-ffpack package with our > system openblas works fine. > the question is why on your system it causes a problem. this setup

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-01-11 Thread Antonio Rojas
openblas is not being built in Sage, the system one is detected and used after FS#27870. FWIW, building our distro fflas-ffpack package with our system openblas works fine. El sábado, 11 de enero de 2020, 10:00:43 (UTC+1), Dima Pasechnik escribió: > > On Sat, Jan 11, 2020 at 8:57 AM Antonio Roja

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-01-11 Thread Dima Pasechnik
On Sat, Jan 11, 2020 at 8:57 AM Antonio Rojas wrote: > > I don't think so - I'm pretty sure I've built it successfully after that > commit. > please compare the system openblas linkage with the one you build in Sage, and post the outputs of ldd here. I am pretty sure there is an interesting diffe

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-01-11 Thread Antonio Rojas
I don't think so - I'm pretty sure I've built it successfully after that commit. El sábado, 11 de enero de 2020, 9:25:04 (UTC+1), Dima Pasechnik escribió: > > isn't it due to https://trac.sagemath.org/ticket/27444 - which removed > --disable-openmp from fflas-ffpack flags ? > see https://trac.

Re: [sage-devel] 9.1.beta0 build failure on Arch

2020-01-11 Thread Dima Pasechnik
isn't it due to https://trac.sagemath.org/ticket/27444 - which removed --disable-openmp from fflas-ffpack flags ? see https://trac.sagemath.org/ticket/27444#comment:34 On Sat, Jan 11, 2020 at 7:53 AM Antonio Rojas wrote: > > Build fails at fflas-ffpack, log attached. Seems related to FS#27870 > >